Computer design of T-cell agonist or antagonist glycopeptides: the effect of sugar identity and anomeric configuration on MHC binding.

The improved chemical and biological properties of synthetic glycopeptides over peptides suggest their use as T cell agonists or antagonists. Recently, we prepared glycopeptide analogues of major T helper cell epitopic peptides corresponding to rabies virus proteins, and experimentally characterized their ability to bind to MHC class II proteins and stimulate T cell clones to rabies virus. In the current study, we investigated these MHC: peptide interactions by molecular modeling. We obtained structural support for our finding concerning the anomeric specificity of MHC with binding. While alpha-linked glycopeptides can bind to MHC without major alterations in the spatial arrangements and hydrogen bonding pattern of class II-peptide binding, the binding of beta-linked glycopeptides is considerably less favorable due to steric and columbic conflicts. Depending on where the saccharides are positioned along the peptide sequence, the MHC: glycopeptide complex may or may not produce the surface profile required for successful T cell receptor interaction. Application of this approach to other antigenic stimuli offers a good model to "dial in" the necessary sugar identity, length and anomeric configuration, as well as promising amino acid mutation sites, for successful design of T cell agonist or antagonist glycopeptides.